1. Diane E. Wickland Earth Science Division NASA HQ NASA Terrestrial Ecology ROSES-2012 Data Product Selection May 1, 2013

2. A.4 ROSES-2012: Terrestrial Ecology The NASA Terrestrial Ecology program requested the following types of research investigations: Data set development in support of Arctic-boreal ecosystem vulnerability research to be conducted in a future Terrestrial Ecology Program-sponsored field campaign (pre- ABoVE) Data set development to meet priority needs of the NASA terrestrial ecological community. Successor studies in the areas of remote sensing science and remote sensing methods development that offer to significantly advance the results of prior NASA Terrestrial Ecology research.

3. A.4: Balance Across Topics (# Selected) Pre-ABoVE (5): Loboda (fire disturbance) Walker (key plot and map data) Munger (historical surface and airborne CO 2 ) Zhang (active layer thickness) Carroll (surface water time series) Successors (3): Lavalle (InSAR temporal decorrelation algorithm) Hanan (tree-grass system LAI/fPAR) Siqueira (PolInSAR and InSAR for forest structure) TE Priority Data Sets (4): Chini (land use transitions data set) Thornton (weather data set) Pelletier (soil depth data set) Townsend (spectral library)

4. A.4 ROSES-2012: Terrestrial Ecology A Two-Step Process was Used:  Step-1 Proposals Due: September 18, 2012  89 Step-1 Proposals Received (2 were late and were not be reviewed)  29 Proposers Encouraged to Submit Step-2 Proposals  Step-2 Proposals Due: January 8, 2013  30 Step-2 Proposals Received  12 Proposals Selected for Funding on April 8, 2013 (solicitation advertised 7-12 selections) Funding advertised: $1.5M/yr for up to 3 years Funding to be awarded: $5.2 million over three years

5. A.4 ROSES-2012: Terrestrial Ecology The NASA Terrestrial Ecology program requested the following types of research investigations: Data set development in support of Arctic-boreal ecosystem vulnerability research to be conducted in a future Terrestrial Ecology Program-sponsored field campaign (pre- ABoVE) Data set development to meet priority needs of the NASA terrestrial ecological community. Successor studies in the areas of remote sensing science and remote sensing methods development that offer to significantly advance the results of prior NASA Terrestrial Ecology research.

6. 1Global re-GAP data set, incorporating LIDAR 2 High spatial and temporal resolution (1-km and 3-hourly) climate, Global coverage. Climate data to include PAR and diurnal variability needed to drive land-surface model/photosynthesis schemes. 3 Gridded species richness and occurrence data to look at biodiversity (at 25-km resolution). Plants and animals (eBird database). (PPBio-INPA in Amazonia) 4High spatial (30-m) and temporal resolution (every 3 to 5 years) Land Cover and Land Cover Change (LandSat based). 5 High spatial resolution remote sensing products for use with cropland data layers (e.g., MODIS at 250-m; crops at 30-m (TM- type) resolution). 1: LandSat used to downscale MODIS. 6 High spatial resolution remote sensing products for use with cropland data layers (e.g., MODIS at 250-m; crops at 30-m (TM- type) resolution). 2: LandSat and Spot combined to obtain higher temporal resolution (seasonal). 7 Hyperspectral Reference Library of leaf-level hyperspectral optical properties (full spectrum; biogeochemical content) and soil reflectivity. Collected using standard measurement techniques; full contextual information. 8Multi-faceted product containing forest height, biomass, and age. (Combining many existing products (LIDAR, LEDAPS). 9 Global database of ground-based in situ validation data for remote sensing products. Forest inventory data (e.g., Height, basal area, biomass, age, crown dimensions of forest study plots globally (validation / evaluation data set). 10 Ground-based data to validate remote sensing products (improved spatial representativeness; match ground-based spatial resolution with that of the satellite data) 11Data on fate of harvested products 12Global gridded data on land-cover, land-use, land-use transitions, and land cover changes (past, present, future) What new data products would advance TE research goals? (1/2)

7. What new data products would advance TE research goals? (2/2) 13 Long-term data records for dynamics of inland waters, cloud corrected global irradiance, atmospheric optical properties, in addition to the existing MEASURES products (landsat, NDVI, etc.). 14High-spatial resolution soil data, including type and depth 15 Global data on soil carbon stocks, litter 16Ancillary data to support interpretation of data from new NASA sensors (e.g., DESDnyI) 17 Gridded products that help us understand land management (not just land cover) are important for substantially improving (a) estimates of carbon fluxes, and (b) reducing uncertainty. 18Atmospheric tracers((N2O, CO) in addition to CO2 for top-down inversion estimates/constraints. 19 N pools and fluxes for crop production, fertilizer use and efficiency, nitrate leaching and N gas emissions, and nitrogen deposition, with a resolution of 0.5 X 0.5 degree at global scale. 20 new MISR product with atmospheric correction and improved ease-of-use (cf MODIS) 21 new LandSat Product with atmospheric correction and georegistration (cf MODIS). Isn't this WELD? 22 Expand spatial and temporal coverage: 1. National Biomass and Carbon Dataset (J. Kellendorfer, WHRC) 23 Expand spatial and temporal coverage: 2. North American ASTER Land Emissivity (S. Hook, JPL) 24 Expand spatial and temporal coverage: 3. North American Forest Dynamics (S. Goward, UMD) 25 Expand spatial and temporal coverage: 4. LEDAPS (J. Masek, GSFC) 26 Expand spatial and temporal coverage: 4. Global Fire Emissions Dataset – GFED (J. Randerson, UCI) 27 Expand spatial and temporal coverage: 5. VULCAN Fossil Fuel Emissions (K. Gurney, Purdue) 28 Data products for evaluating and improving model performance (I-LAMB parameters) LAI, NPP, CO2 annual cycle (phase & amplitude), energy & CO2 flux

8. 11/12/12 Hi Diane, A group of us have put together a description of priority new data sets that will advance NASA TE goals. At the TE Team meeting you said you'd like this information by mid-November. The group that pulled this information together includes George Hurtt and myself, plus Peter Thornton, Dan Hayes, Joanne Nightingale, Mac Post, Simon Hook, and Fred Hummerich. We tried to focus on the rationale that makes the data a high priority and the characteristics of the data. The description of how the data might be compiled was left out of the description. I wanted to pass this along to you now as a DRAFT, to get your input. Is this what you are looking for? How can this be improved? It is probably too long, but we hoped that it described the data sufficiently. Does this document need an introduction? Does the text need a statement like this?: "Efforts to address these various data products must be handled with strong attention to issues of data consistency, service, management and usability, which will require close coordination between the dataset developers, model user communities, and service providers." I am still waiting for a description for observations that would benefit model evaluation and benchmarking activities (ILAMB). A group of us (MAST-DC, Michalak's MsTMIP, ILAMB, DataONE, and some NSF and DOE projects) is having a major model intercomparison workshop next week at which time we'll discuss those ILAMB observations. Best wishes, Bob ROSES-2012 A.4: Priority Data Sets for Terrestrial Ecology

9. Data needed to advance Terrestrial Ecology Research Goals 1. High spatial and temporal resolution weather data One of the fundamental forcings for land process models is surface weather at the land-atmosphere boundary. As land process models move toward higher spatial and temporal resolution and larger spatial domains, the resolution requirements for surface weather boundary conditions also increase. Over the coming decade, land process models will operate at sub-kilometer spatial and sub-hourly temporal resolution for global domains, forced in part by land surface state and process information (e.g. soil type and disturbance history) at sub- kilometer resolution. A new generation of surface weather boundary conditions is required to meet the needs of these models, with sub- kilometer resolution, global land coverage, and on the order of hourly temporal resolution. These datasets should include near-surface air temperature, precipitation, humidity, incident short and longwave radiation, and winds. Because land processes are known to be sensitive to variability in weather and climate on temporal scales ranging at least from hours to centuries, these surface weather boundary conditions should be provided for the longest possible historical period, and updated at least annually with the latest observations. Methods must be developed to provide the most robust estimates of surface weather in the face of low and variable observation density, and the estimation accuracy of these surface weather datasets must be quantified to the fully possible extent, providing a foundational component for land process model uncertainty quantification. 2. Global gridded data on land-cover, land-use, land-use transitions, and land cover changes (past, present, future). The impacts of disturbance and land cover / land use change have long been recognized to be a key driver of climate and ecosystem function at local to global scales, via feedbacks through carbon, nutrient, water and energy cycles. Despite this, the incorporation of these impacts in land surface models – needed for reliable retrospective and predictive assessments of climate variability within coupled Earth System Model frameworks – remains an on-going challenge. Models have been developed to prescribe disturbance mechanisms, but progress in this area has been limited. Needed are consistent and comprehensive driver data sets on disturbance and land use change available at the spatial and temporal scales necessary for models to represent their impacts on land-atmosphere interactions. In developing these new data sets, it should be recognized that a comprehensive set of disturbance forcings from various data sources need to be organized using a consistent approach that can be readily incorporated into model systems (i.e. a gridded and temporally explicit framework) along with the contributing change agent. The legacy of past disturbances (on the scale of decades to centuries) defines contemporary state variables and ecosystem function, but here we are limited by the lack of data prior to the start of historical data sets. While these data sets can and should be informed by satellite-era data, more complete historical disturbance data sets covering the temporal scales needed for model “spin-up” (centuries to millennia) will require integration with new tools and data sources. Furthermore, to improve climate forecasts, scenarios for future disturbance and land use change are also a high priority. ROSES-2012 A.4: Priority Data Sets for Terrestrial Ecology

10. Data needed to advance Terrestrial Ecology Research Goals (cont.) 3. In situ validation requirements for Terrestrial Essential Climate Variables In support of the United Nations Framework Convention on Climate Change (UNFCCC), the Global Climate Observing System (GCOS) has specified the need to systematically produce and validate a suite of key global satellite-derived parameters that play an important role in understanding land-surface - climate interactions. Among the 50 Essential Climate Variables (ECVs) listed, those relevant to the land remote sensing community include: snow cover, permafrost, land cover, fire disturbance, albedo, leaf area index (LAI), the fraction of absorbed photosynthetically active radiation (*APAR), above ground biomass, and soil moisture (hxxp://gosic.org/ios/MATRICES/ECV/ecv- matrix.htm). In the terrestrial domain, it is necessary to obtain global products for most ECVs from a range of satellite sensors supported by in situ measurements. A set of in situ terrestrial observations at a diversity of sites is needed for: (a) validation of observations derived from Earth observation satellites; (b) process studies; (c) observations of the fullest possible range of terrestrial ECVs and associated details relevant to their application in model validation; and (d) to address intrinsic limitations in derived products. In the context of long-term global climate measurements, observations are critically needed that will provide consistent in situ measurements for product validation activities at reference sites within diverse biomes, adding to and leveraging from existing core validation networks.hxxp://gosic.org/ios/MATRICES/ECV/ecv- matrix.htm 4. High spatial resolution soil depth data layer. A major difficulty with current soils data used for terrestrial biogeochemistry modeling is the representation of soil depth. The current depth limit of 1-m is an insufficient depth for ecohydrological analysis and evaluation for many natural terrestrial systems. While the next generation Global high-resolution soils data (GlobalSoilMap.net) is being compiled, it won’t be completed for another 5 to 10 years. In the interim novel methods to develop high spatial resolution data on soil depth regardless of the distance to bedrock must be developed to advance Terrestrial Ecology modeling. ROSES-2012 A.4: Priority Data Sets for Terrestrial Ecology

11. Data needed to advance Terrestrial Ecology Research Goals (cont.) 5. Spectral Libraries The value of airborne and spaceborne datasets covering the spectral range 0.35-13 um is enhanced with ground measurements, typically made at higher spectral resolution of different vegetation types. In order to maximize the usefulness of these ground measurements, they need to be made under known and documented conditions with well-calibrated instrumentation and made freely available with standardized format and content. These ground datasets are typically referred to as spectral libraries. Unfortunately there are very few spectral libraries of vegetation and the measurements in them typically only cover a limited spectral range or a few vegetation types. Perhaps the most difficult aspect to making a measurement is the need to measure the vegetation over the range of conditions that it typically experiences, e.g. different soils and environmental conditions as well as the behavior of the full canopy and changes associated with the phenological cycle. The Terrestrial Ecology program solicits proposals to compile existing measurements and contextual information and make additional measurements, focusing on laboratory measurements before more field measurements are made. It is important that the measurements made are carefully documented, cover the full optical range used by the airborne and spaceborne data, and are made at higher spectral resolutions. Observations compiled in the library may include: leaf level observations of spectral reflectance and transmittance as inputs to radiative transfer models, non-green material (e.g., bark, branches, dead leaves, leaf litter), and other plant material (e.g., lichens, mosses, etc.). Ancillary data on measurement context should also be part of the data record (species, growth stage, leaf canopy position, location and date, biochemical measurements (e.g. chlorophyll content, nitrogen content, etc.), and photosynthetic response parameters (Amax, quantum efficiency) Information needs to robust enough to examine between and within species variability. ROSES-2012 A.4: Priority Data Sets for Terrestrial Ecology

12. The five top-priority data sets solicited were:  High spatial and temporal resolution weather data  Global gridded data on land-cover, land-use, land-use transitions, and land-cover changes (past, present, future)  In situ validation data sets for Terrestrial Essential Climate Variables (No Step 2 Proposals Encouraged or Received)  High spatial resolution soil depth data layer  Spectral libraries ROSES-2012 A.4: Priority Data Sets for Terrestrial Ecology

13. A.4 ROSES-2012: Terrestrial Ecology The four top-priority TE data sets selected were: Surface weather data with uncertainty quantification for terrestrial ecosystem process models (Thornton, Peter / Oak Ridge National Laboratory) Using NASA Remote Sensing Data to Reduce Uncertainty of Land-use Transitions in Global Carbon-Climate Models (Chini, Louise / University of Maryland, College Park ) Development of a High-Resolution Global Soil Depth Dataset (Pelletier, Jon / University Of Arizona) Ecological Spectral Information System (ESIS): Integration of Spectral Data with Measurements of Vegetation Functional Traits (Townsend, Philip / University of Wisconsin, Madison)